This work proposes a practical methodological framework for simulating long-term discrete wind speed time series, for applications to renewable energy systems. The framework is based on the Matérn stochastic process, that is used to identify and extract the short and long-term periodic components present in the mean and variance of a wind speed time series, and to characterize the residual random component. The procedure of construction of synthetic wind speed time series is developed, rooted in a fast numerical simulation algorithm for the Matérn process. We empirically validate the usefulness of the proposed framework by numerical experiments based on three real historical wind speed datasets. The results show consistent statistical similarities between the historical and simulated wind speed time series of data.